Skip to main content

A user found the bug I should have caught myself

I was collecting user feedback on Iqra a few weeks ago. One of the users pointed out a bug. A section of the app was not working correctly. I knew the section. I had made changes to it recently — and shipped them directly to production without running through my usual testing flow.

A user told me about the bug. Not my emulator. Not my test release. A user.

I fixed the bug, made a few more improvements, added some things to the app. When I finally sat down to build the test release properly, this came back to me. So here it is — written down, for anyone building a product right now.

The flow that actually protects you

Every change, no matter how small, goes through three stages before it reaches your users.

First, test on the emulator. Catch the obvious breaks early, before anything leaves your machine. Second, publish a test release and run it on a real device. The emulator lies sometimes — screen sizes behave differently, performance feels different, interactions that work in simulation break on actual hardware. Third, only then push to production.

It takes more time than shipping directly. It has saved me more time than it has cost me.

Why developers skip this

It is not always laziness. Sometimes it is a small change — one line, one fix — and it feels unnecessary to run the full flow for something that minor. That feeling is exactly when the mistake happens. The smaller the change feels, the easier it is to justify skipping the test. And the skipped test is always the one that finds you later, through a user report.

The honest note

If you already do this — good. This was not written to teach you something you know. It was written because someone reading this is about to push a change directly to production right now, and this might be the thing that stops them.

Your users should not be your test environment.

Labels:
Location: Lahore, Pakistan

Comments

  1. AnonymousApril 23, 2026 at 10:08 PM

    Thanks for sharing your experiences. Everyone says, learn and earn. No body talks about the stuff in between.

    ReplyDelete

Post a Comment

Share your thoughts

Popular posts from this blog

Earning after the advent of AI, the 3 paths and things that matter

I have received this question more times than I can count. Are jobs gone? Is AI taking over everything? Why are internships not coming through? What do I do with my degree or my skill? Here is my honest answer — not a motivational one. AI has made things easier, and harder at the same time. Because AI can now do multiple things, the person who knows how to get the right things done from AI is preferred over the one who does not. That is the new baseline. Learn the basics of AI regardless of which path you choose — not to become an AI engineer, but to use it well in whatever you are doing. Beyond that, I see three paths. Each one has its own direction. The mistake most people make is not choosing one and going deep on it. Path 1: Build a Personal Product This is the path I am on. I will be honest about where my experience ends. Start by finding a problem — ideally one you personally face. That is your built-in validation. If you have the problem, others likely do too....
Post a Comment
Read more

How an AI Student is Earning from LinkedIn — and What I Learned from a 17-Minute Call

A few days ago I saw a post on Facebook. Someone sharing that he had started learning AI, done some projects, and was now earning from it. I asked him how. He said he got a client through LinkedIn. I asked him to elaborate. He said let's talk. Seventeen minutes on WhatsApp. That was the whole conversation. But it connected more dots than months of reading about "the opportunity in AI." The problem with how most people learn AI AI is a wide field. Machine learning. Deep learning. Model development. Model deployment. Data analysis. Data analytics. Automation. Most people learning AI are aware of all of these — and that awareness becomes the problem. They keep exploring, keep switching, keep adding things to the list. A year passes and they are still learning. He told me something simple: to stand out and to get clients, you have to focus on one thing. His one thing is RAG — Retrieval Augmented Generation. Not AI in general. Not even a broad category within AI...
Post a Comment
Read more